public class DFunctionBatchNorm_F64 extends BaseDBatchNorm_F64 implements DFunctionBatchNorm<deepboof.tensors.Tensor_F64>
DFunctionBatchNorm for Tensor_F64. Intermediate variables are cached in the
forward pass.D, EPS, params, requiresGammaBeta, shapeVariables, tensorDiffX, tensorDMean, tensorDVar, tensorDXhat, tensorMean, tensorStd, tensorTmp, tensorXhatlearningMode| Constructor and Description |
|---|
DFunctionBatchNorm_F64(boolean requiresGammaBeta) |
| Modifier and Type | Method and Description |
|---|---|
protected void |
_backwards(deepboof.tensors.Tensor_F64 input,
deepboof.tensors.Tensor_F64 dout,
deepboof.tensors.Tensor_F64 gradientInput,
java.util.List<deepboof.tensors.Tensor_F64> gradientParameters) |
void |
_forward(deepboof.tensors.Tensor_F64 input,
deepboof.tensors.Tensor_F64 output) |
protected int[] |
createShapeVariables(int[] shapeInput)
Create the shape for all the variables which are being normalized
|
void |
forwardsEvaluate(deepboof.tensors.Tensor_F64 input,
deepboof.tensors.Tensor_F64 output) |
_initialize, _setParameters, getEPS, getMean, getTensorType, getVariance, hasGammaBeta, setEPSbackwards, evaluating, isLearning, learningforward, getOutputShape, getParameters, getParameterShapes, initialize, setParametersclone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, waitforward, setParametersgetMean, getVarianceprotected int[] createShapeVariables(int[] shapeInput)
BaseDBatchNorm_F64createShapeVariables in class BaseDBatchNorm_F64shapeInput - Shape of input tensor (without mini-batch)public void _forward(deepboof.tensors.Tensor_F64 input,
deepboof.tensors.Tensor_F64 output)
_forward in class deepboof.impl.forward.standard.BaseFunction<deepboof.tensors.Tensor_F64>public void forwardsEvaluate(deepboof.tensors.Tensor_F64 input,
deepboof.tensors.Tensor_F64 output)
protected void _backwards(deepboof.tensors.Tensor_F64 input,
deepboof.tensors.Tensor_F64 dout,
deepboof.tensors.Tensor_F64 gradientInput,
java.util.List<deepboof.tensors.Tensor_F64> gradientParameters)
_backwards in class BaseDFunction<deepboof.tensors.Tensor_F64>